Re: [soa-rm-ra] OO vs. SOA and what goes into a service

[Danny Thornton <danny_thornton2@yahoo.com>]

The message names can denote the action to take -
AddResourceRequest and AddResourceResponse for
example.  If CRUD operations are performed by a
service, the actions can be realized through CRUD
message types and could be handled by a single
service.  You can define real world effects or
constraints for the action/message just like any other
action/message the service may receive.

In the above model, the interface can have a single
method (like receiveMessage) where the received
message type denotes the action.

The other model is to have separately named methods in
the service for the actions.

A real difference for a service and an OO object is
that service functionality is self contained and it
does not flow across the wire, only the information in
the information model flows across the wire.  With an
OO object both the data and the functionality of the
object flow across the wire.  This is why OO objects
with functionality reak havoc when they are
distributed.  For example, I may accept some
distributed OO object and then find out I have to
compile in another 5,000 libraries (that's how it felt
sometimes) to use it because of all the functionality
that came with it.  

Danny

--- Ken Laskey <klaskey@mitre.org> wrote:

> Is it fair (or at least not too distorted) to say
> that with OO we  
> define an object and look for what we can do to it
> (i.e. its methods)  
> while with SOA we identify what we want to do (i.e.
> business  
> functions) and then, if appropriate, look for
> objects to do it to?
> 
> This gets back to the discussion last October on
> what actions get  
> bundled together in a service.  From an OO
> perspective, I choose an  
> object and then attach CRUD methods.  In SOA, would
> the CRUD methods  
> individually be things I want to do, each with a
> describable real  
> world effect (and possibly policies on who can do it
> and under what  
> conditions), and I define services to carry out
> those functions?
> 
> Are both perspectives/design approaches equally
> valid?  What are the  
> implications for description and discovery?
> 
> Ken
> 
>
------------------------------------------------------------------------
> 
> ------------------
> Ken Laskey
> MITRE Corporation, M/S H305     phone:  703-983-7934
> 7515 Colshire Drive                        fax:     
>   703-983-1379
> McLean VA 22102-7508
> 
> 



      ____________________________________________________________________________________
Be a better friend, newshound, and 
know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ